Purpose. To compare the performances of three evaluation methods in detecting deficiencies of professionalism among third-year medical students during their ambulatory care and inpatient ward rotations of a core internal medicine clerkship. Method. From 1994 to 1997, 18 students at The Uniformed Services University of the Health Sciences failed to satisfactorily complete their core 12-week third-year internal medicine clerkship due to deficiencies in professionalism. Three evaluation methods had been used to assess all students' professionalism during the two rotations of their clerkship: standard checklists, written comments, and comments from formal evaluation sessions. Using qualitative methods and the information obtained by the three evaluation methods, the authors abstracted the record of each student concerning his or her clerkship behavior in terms of the six domains of professionalism used on the standard checklist. A detection index, which is the percentage of all instructors' less-than-acceptable ratings of a student across the six professionalism domains, was calculated for each evaluation method for each of the two clerkship settings. Results. For each evaluation method, deficiencies in professionalism were twice as likely to be identified during the ward rotation as during the ambulatory care rotation (p < .002 for all). Formal evaluation session comments had the highest detection index in both clinical settings. Although the numbers of written and formal evaluation session comments per evaluator and per cited professionalism domain were similar, nearly a fourth of the instructors made identifying comments at the evaluation sessions only. Conclusion. In the clerkship studied, deficiencies in professionalism of such magnitude as to require remediation were more likely to be identified in the inpatient than in the ambulatory care setting. Of the three evaluation methods studied, the face-to-face, formal evaluation sessions significantly improved the detection of unprofessional behavior in both clerkship settings. Further efforts at such an interactive evaluation process with ambulatory care clerkship instructors may be essential for improving the identification of unprofessional behavior in that setting.